Response surface methodology is being developed to analyze the results of combination chemotherapy experiments in mice bearing L1210 or P388 leukemia. Several different models of increasing complexity have now been successfully utilized including the logistic, a proportional hazard model and a non-proportional hazard model. Independent variables including drug dose, length of time between administration of drugs and age of the treated tumor are being studied where the response variable is time to death. These studies not only provide more detailed information about the drugs used and their interactions but can serve as a guide for possible application of these methods to clinical combination chemotherapy or combined modality treatments. In that regard simulation studies are being undertaken to determine efficient design criteria including how many groups and how many individuals in each group are required to give results utilizing the minimum number of individuals for the entire investigation.